HyperTransport in Server Applications

HyperTransport Enables Servers to Deliver Top Performance

High performance servers are mission critical components of the enterprise network, delivering web pages, enterprise data and middleware applications to corporate users across internal and external networks. Often using multiple processors, these systems are optimized to deliver maximum data transfers between legacy I/O subsystems and high bandwidth network communications technologies.

HyperTransport-enabled processors incorporate a low-latency link between CPU and CPU as well as a means of integrating multiple I/O streams in an efficient, low overhead mechanism that is compatible with high bandwidth technologies such as Gigabit Ethernet and with legacy I/O systems such as PCI and PCI-X interfaces. With HyperTransport technology's inherent ability to map to PCI Express, such systems can easily incorporate future developments using that technology as well.

The key to HyperTransport in the server is that HyperTransport delivers a low latency means to simplify and integrate multiple I/O subsystems in a cost and printed circuit board-efficient manner.  As shown in the diagram, HyperTransport can be used a primary link between the core processors and with HyperTransport bridges and hubs to link to PCI, PCI-X and PCI Express I/O subsystems.